Chapter 5

Introduction to Advanced Macroeconomic

(Scarth, 2009)

The Challenge of
New Classical
Macroeconomics
Abdul Hadi Ilman
Universitas Teknologi Sumbawa
November 2014

Introduction
in Chapter 5, we examine the New Classical approach to business cycle analysis the modern,
more micro-based version of the market-clearing approach to macroeconomics, in which no
appeal to sticky prices is involved
Our analysis thus far may have left the impression that all macroeconomists feel comfortable with
models which "explain" short-run business cycles by appealing to some form of nominal rigidity.
The Pioneer of new synthesis would regard some of underlying assumption concerning of the
sticky price as heroic
New Keynesians reaction is to explain to foundation of menu cost (the cost of changing nominal
prices), and to elaborate how other features, such as real rigidities and strategic
complementarities, make the basing of business cycle theory on seemingly small menu costs
appealing after all
New Classicals reaction to the proposition that menu costs "seem" too small to explain business
cycles is to investigate whether cycles can be explained without any reference to nominal
rigidities at all.

Introduction
The chapter is organized as follow:
1. The original real business cycle model
2. Extension of the basic model
3. Optimal inflation policy
4. Harberger tiangles vs Okuns Gap

Part 1

The Original Real Business

Cycle Model

Keynesian analysis explains the business cycle by assuming rigid money wages.
Demand shocks push the price level up in booms and down in recessions. This set of
outcomes makes the real wage rise in recessions and fall in booms, and the resulting
variations in employment follow from the fact that firms slide back and forth along a given
labour demand curve. Thus, the model predicts that the real wage moves contracyclically
(and this is not observed), and because the labour market is not clearing, it makes the
labour supply curve irrelevant for determining the level of employment.
New Classicals prefer to assume that wages are flexible and that the labour
market always clears. To their critics, a model which assumes no involuntary
unemployment is an "obviously" bad idea. But if this is so, say the New Classicals, it should
be easy to reject their theory. While the approach has encountered some difficulties when
comparing its predictions to the data, it has turned out to be much more difficult to reject
this modelling strategy than had been first anticipated by Keynesians.

Part 1

The Original Real Business

Cycle Model

New Classicals explain business cycles by referring to real shocks (shifts in

technology) so the approach is called real business cycle analysis. The basic idea is that
workers make a choice concerning the best time to work.
If something happens to make working today more valuable (such as an increase in
today's wage compared to tomorrow's) workers make an inter-temporal substitution; they
work more today and take a longer than usual vacation tomorrow. Similarly, an increase
in interest rates lowers the present value of future wages, and so leads individuals to
work more today.
If there is a positive technology shock today, the higher marginal product of labour implies a
shift to the right of the labour demand curve. Similarly, a negative development in the
technology field next period shifts the labour demand curve back to the left. Thus, according
to this view, business cycles are interpreted as shifts in the position of the labour
demand curve, not movements along a fixed labour demand curve.

Part 1

The Original Real Business

Cycle Model

Hansen and Wright's (1992) version real business cycle model

Household utility
function
Time constraint
Production function
Technology shock
Accumulation of capital
stock

Part 1

The Original Real Business

Cycle Model

Equation (5.1) is the household utility function; E, , C and L denote expectations,

the rate of time preference, consumption, and leisure.
Equation (5.2) is the time constraint (N is employment). Formal utility maximization
is used to derive consumption and labour supply functions.
Equations (5.3), (5.4) and (5.5) define the production function and the one stochastic
variable in the system the technology shock, z. An ongoing trend is involved, and
the stochastic part is a normally distributed error term with zero mean and constant
variance, and is the coefficient of serial correlation in this exogenous process.
Formal profit maximization is used to derive the demands for capital and labour.
Finally, equations (5.6) and (5.7) define the accumulation of the capital stock is its
depreciation rate and I is gross investment)

Part 1

The Original Real Business

Cycle Model

If the model's data "looked like" the real world data, researchers concluded that this simple
approach has been vindicated.
The result showed stylized facts of the business cycle are well illustrated by the "data" that
is generated from this very simple structure
Diminishing marginal utility of consumption
consumption is less variable than income,
investment is more volatile than income

If a positive technology shock occurs today, people can achieve higher utility by spreading
out this benefit over time. So they increase consumption by less than their income has
increased initially, and as a result, some of the new output goes into capital accumulatio
It is impressive that these very simple calibrated models can mimic the actual magnitude
of investment's higher volatility relative to consumption

Part 1

The Original Real Business

Cycle Model

Despite these encouraging results, however, this model does not

generate the wide variations in employment and the very low
variability in real wages (that we observe in real data) without an
implausibly high value for the wage elasticity of labour supply (an
unacceptable value for parameter ).

Part 2

Extensions to the Basic Model

We are considering a number of extensions to the basic New Classical model that
have been offered as mechanisms that can make the "data" that is generated from
the calibrated models more representative of the actual co-movements in real
wages and employment
The first extension we consider is an alteration in the utility function. The one given
in equation (5.1) embodies the assumption that the marginal utility of leisure in one
period is not affected by the amount of leisure enjoyed in other periods
The modern approach to adapting the utility function (central to much recent New
Classical work) is to specify that today's utility depends on both today's level of
consumption and today's level of habits. Habits evolve over time according to the
following relationship:

Part 2

Extensions to the Basic Model

Another extensions:
Indivisible labour. At the macro level, then, variation in employment comes from
changes in the number of people working, not from variations in average hours per
worker. This means that the macro correlations are not pinned down by needing to be
consistent with evidence from micro studies of the hours supplied by each individual
(that show a very small elasticity).
Non-market activity. "home" production is a very significant amount of real economic
activity. Benhabib, Rogerson, and Wright (1991) have shown that when the real business
cycle model involves this additional margin of adjustment, its real wage-employment
correlations are much more realistic. The basic idea is that the amount of leisure
consumed can remain quite stable over the cycle even while measured employment
in the market sector of the economy is changing quite dramatically when households
have the additional option of working at home (doing chores for which they would
otherwise have paid others to do).

Part 2

Extensions to the Basic Model

Variations in government spending. This work involves adding an additional
(demand) shock (in addition to technology shocks) to the model; for example,
is added to the system, and the market clearing condition is changed to
Increases in government spending raise interest rates, and higher interest rates
decrease the present value of working in the future. With the relative return of
working in the current period thereby increased, the current-period labour supply
curve shifts to the right. With these supply-side shifts in the model, some of the
variations in employment are explained by shifts along a given labour demand
curve. With this additional source- of employment fluctuation, the magnitude of the
technology shocks does not need to be as great for the calibrated model to
generate realistic changes in employment.

Part 2

Extensions to the Basic Model

The variability of price mark-ups over the cycle is an example of a
demand-shift mechanism, we know that the mark-up of price over
marginal cost falls during booms because of the entry of new firms. This
fact causes the labour demand curve to shift to the right during booms.
Households shifting between market-oriented employment and
home production is an example of a supply-shift mechanism.
Payment lag. If firms have to pay their wage bill one period before
receiving their sales revenue, the labour demand function becomes FN =
(W/P)(1+r) so variations in the interest rate shift the position of the
labour demand curve.

Part 2

Extensions to the Basic Model

It is interesting to note the convergence involved with parts of New Keynesian and
New Classical work.
Keynesians have been taking expectations and micro foundations more seriously to
improve the logical consistency of their systems,
while Classicals are embracing such things as autonomous expenditure variation,
imperfect competition and payment lags to improve the empirical success of their
models.
Despite this convergence, however, there is still a noticeable difference in emphasis.
Classical models have the property that the observed fluctuations in employment
have been chosen by agents, so there is no obvious role for government to reduce
output variation below what agents have already determined to be optimal. This
presumption of social optimality is inappropriate, however, if markets fail for any
reason (such as externalities, moral hazard, or imperfect competition).

Part 2

Extensions to the Basic Model

Hansen and Wright (1992) have shown that when all of these
extensions are combined, a simple aggregative model can
generate data that reflects fairly well the main features of the
real-world real wageemployment correlations after all.
But still the model does not fit the facts well enough, so even
pioneers of this approach (for example, Goodfriend and King
(1997) have called for the New Neoclassical Synthesis in which
temporarily sticky prices are added to the real business cycle
model. We consider this synthesis in some detail in Chapter 6

Part 2

Extensions to the Basic Model

It may seem surprising that New Classicals have embraced the hallmark of
Keynesian analysis sticky prices. Why has this happened?
Perhaps because there is one fact that appears to support the relevance of
nominal rigidities the well-documented correlation between changes in the
nominal money supply and variations in real output. Either this is evidence in
favour of nominal rigidities, or it is evidence that the central bank always
accommodates increasing the money supply whenever more is wanted
(during an upswing)
One final consideration is that real and nominal exchange rates are very highly
correlated. Many economists argue that there appears to be no way to account
for this fact other than by embracing short-run nominal rigidities.

Part 2

Extensions to the Basic Model

Keynesians welcome this convergence of research approaches, yet (at the conceptual
level) they remain concerned about the lack of market failure involved in the classical
tradition. Also, they have some empirical concerns:
The real business cycle approach is based on the notion of intertemporal substitution of
labour supply. However, micro studies of household behaviour suggest that leisure and the
consumption of goods are complements, not substitutes (as assumed in New Classical
theory).
Another awkward fact is that, in the United States at least, only 15 percent of actual labour
market separations are quits. The rest of separations are layoffs. In addition, the data on
quits indicates that they are higher in booms. The real business cycle model predicts that all
separations are quits and that they are higher in recessions.
Finally, it is a fact that a high proportion of unemployment involves individuals who have
been out of work for a long time an outcome that does not seem consistent with the
assumption of random separations.
It is impossible to observe the technology shocks directly.

Part 2

Extensions to the Basic Model

Given these problems, why does real business-cycle theory appeal to
many of the best young minds of the profession?
Blinder (1987) attributes the attraction to "Lucas's keen intellect and profound
influence," but it also comes from the theory's firm basis in microeconomic
principles and its ability to match significant features of real-.world business
cycles
Given these problems, why does real business-cycle theory appeal to many of
the best young minds of the profession? Blinder (1987) attributes the attraction
to "Lucas's keen intellect and profound influence," but it also comes from the
theory's firm basis in microeconomic principles and its ability to match
significant features of real-world business cycles

Part 3

Optimal Inflation Policy

To justify a zero-inflation target, many analysts make the following argument.
Since inflation is a tax on the holders of money, and since taxes create a loss of consumer surplus
known as an "excess burden" that tax (inflation) should be eliminated.
The problem with this argument is that if one tax is eliminated, the government must raise another
tax (and that other tax creates an excess burden of its own).
Recognizing this, the standard approach in public finance is to recommend the "inverse elasticity
rule" for setting taxes.
Since excess burdens are bigger when taxes cause large substitution effects, the efficiency criterion
for judging taxes stipulates that the largest tax rates should apply to the items that have the
smallest price elasticities of demand.
Since the interest rate is the (opportunity) cost of holding money, and since the estimated interest
elasticity of money demand is very small, the inverse elasticity rule can be used to defend a
relatively large tax on money.
In other words, it supports choosing an inflation rate that is (perhaps well) above zero.

Part 3

Optimal Inflation Policy

The full-equilibrium benefits of low inflation are independent of
the complexities of the transition path that the economy takes to
reach the long run (such as those caused by nominal rigidities). As
a result, all analysts agree that a basic version of the New
Classical model is the appropriate vehicle to use for estimating
the benefits of that policy.
Since our focus is on the longterm benefits of low inflation, we use
an example of New Classical work that highlights money
Mansoorian and Mohsin (2004).

Part 3

Optimal Inflation Policy

Households maximize a standard utility function. As usual,
instantaneous utility is a weighted average of leisure and
consumption
U = ( In(1 - N) + (1 - ) ln C),
and there is a constant rate of time preference, .
The budget constraint is
Consumption is the sum of wage and employment income, plus
the transfer payments received from the government, , minus
the inflation tax incurred by holding real money balances and
minus asset accumulation.

Part 3

Optimal Inflation Policy

The full-equilibrium version of the model is described by the following equations:

The first equation is the Ramsey consumption function when consumption growth is zero (that
follows from the household differentiating with respect to variable C).
The second equation is what emerges when the labour supply function (that follows from the
household differentiating with respect to variable N) is equated with the firms' labour demand
function.
The third equation is the production function, and
the fourth is the other relationship that follows from profit maximization that capital is hired
to the point that its marginal product equals the interest rate.
The final three equations can be used to determine how consumption, employment and the
capital stock respond to different inflation rates.

Part 3

Optimal Inflation Policy

As noted, the government budget constraint is
This equation states that lump-sum transfer payments, T, are paid to
individuals, and in aggregate, these transfers are financed by the
inflation tax
With endogenous, cutting inflation is unambiguously "good". Lower
inflation eliminates tax that distorts the household saving decision, and
no other distortion is introduced by the government having to levy some
other tax to acquire the missing revenue. Consumption, employment,
output and the capital stock all increase by the same percentage when
the inflation rate is reduced. Specifically,

Part 3

Optimal Inflation Policy

Mansoorian and Mohsin calibrate the model with standard real-businesscycle
assumptions: = 0.64, = 0.042, = 0.30 and they assume an initial
inflation rate of zero: = 0. These assumptions allow them to evaluate the
inflation multiplier. The result is that creating inflation of 2% lowers steadystate consumption by 1.37%. This outcome is an annual annuity.
With no growth and a discount rate of 0.042, the present value of this annual
loss in consumption is 0.0137/0.042 = 32% of one year's level of
consumption.
Most analysts regard this magnitude as quite large. The policy implication is
that even a two-percent inflation rate should not be tolerated.

Part 3

Optimal Inflation Policy

Experience has shown (see Ball (1994)) that we suffer an increase in the GDP
gap of about two percentage points for about 3.5 years to lower the inflation
rate by two percentage points. This means that we lose approximately 7
percentage points of one year's GDP to lower steady-state inflation by this
amount. The benefit-cost analysis says that the present value of the benefits
exceeds this present cost (32% exceeds 7%), so disinflation is supported.
This is the standard defense for targeting zero inflation. It is an application of
the neoclassical synthesis. The long-run benefits of lower inflation are
estimated by appealing to our theory of the natural rate (New Classical
macroeconomics). The short-run costs are estimated by appealing to a more
Keynesian model of temporary deviations from that natural rate that stem
from temporary nominal rigidities

Part 3

Optimal Inflation Policy

In diagrammatic terms, the
reasoning is illustrated in Figure 5.1,
where we continue to abstract from
any ongoing growth. New Classical
analysis is used to estimate the shift
up in the potential GDP line that
accompanies disinflation, and the
short-run synthesis model (in which
potential GDP is exogenous) is used
to calculate the temporary drop
(and then later recovery) in actual
GDP.

Part 3

Optimal Inflation Policy

One unappealing aspect of this estimate of the long-term benefits
of low inflation is that it involves the assumption that the
monetary authority can dictate to the fiscal policy maker (and
insist that the latter must cut transfer payments in the face of
disinflation).
How is the analysis affected if we assume that this is not possible?
To explore this question, we add a tax on wage income. Assuming
(as we have already) that the wage equals the marginal product
of labour, we re-express the government budget constraint as

Part 3

Optimal Inflation Policy

There is one change in household budget constraint, since it must now stipulate that
households receive only the after-tax wage. This results in only one change; the third
equation in our list of the model's relationships becomes:
It is more difficult for disinflation to be supported in this case, since one distortion (the
wage-tax) is replacing another (the inflation tax). Indeed, in this case, for the calibration
assumed by Mansoorian and Mohsin, it turns out that the "benefit" of lower inflation has
to be negative!
Disinflation forces the fiscal authority to rely more heavily on a more distortionary
revenue source than the inflation tax. Similar results in more elaborate calibrated
models are reported in Cooley and Hansen (1991). It would therefore appear that the
public-finance approach (that simplifies by using a New Classical model with no ongoing
growth) does not lead to a solid under-pinning for a zero inflation target.

Part 3

Optimal Inflation Policy

A somewhat more reliable argument for choosing a very low inflation
rate (perhaps zero) as "best," concerns the effect of inflation on savings
in a growth context. Most tax systems are not fully indexed for inflation.
To appreciate why this is important, suppose you have a $100 bond that
gives you a nominal return of 10%. Suppose that inflation is 5%, and
that the interest rate on your bond would be 5% if inflation were zero.
At the end of the year, the financial institution sends you a tax form
indicating that you received $10 of interest earnings, and the
government taxes you on the entire $10. In fact, however, only $5 of
the $10 is interest earnings. The other $5 is compensation for the fact
that the principal value of your investment has shrunk with inflation

Part 3

Optimal Inflation Policy

An interest income tax system should tax only interest income,
not the saver's depreciation expenses. An indexed tax system
would do just this.
Thus, inflation reduces the incentive to save, so that individuals
living in the future inherit either a smaller capital stock (with
which to work) or a larger foreign debt to service, or both.
According to this analysis, then, to avoid a lowering of future
living standards, we should pursue a "zero" inflation target.
The case for zero inflation is far from complete

Part 4

Harberger Triangles vs. Okun's

Gap
Much

of this book has focused on

explanations of the business cycle
and an evaluation of stabilization
policy. It has been implicit that
there would be significant gains
for society if the business cycle
could be eliminated
Without business cycles, actual
output, y, would coincide with the
natural rate, ; , and both series
would follow a smooth growth
path such as the straight line